Feature Story

Lightening the cancer protocols’data demands

September 2002
Anne Paxton

The College’s protocols for cancer reporting were never intended to be marching orders. Developed and refined over the last 15 years, the 38 sets of guidelines have been disseminated as educational tools. But soon, as the College approves revised versions of the protocols, they will no longer be strictly voluntary. The essential data elements they specify will become required parts of pathology reports on cancer specimens at the nation’s more than 1,400 approved cancer programs, as the protocols become a new component of the American College of Surgeons’ Commission on Cancer accreditation process.

The Commission on Cancer’s decision in 2000 to make the protocols mandatory has rattled and even angered some members of the pathology community, who maintain that the protocols weren’t designed for this use, that many of the data elements they include are secondary or trivial, and that converting pathology reporting systems in some cases could be costly.

But CAP officials see no cause for alarm. They emphasize they are narrowing down the number of required elements and seeking a longer time frame to make the transition as smooth and unrushed as possible. The Commission on Cancer, or CoC, had planned to make the protocols mandatory as of Jan. 1, 2003, but at press time, the CAP’s leaders had decided to request a one-year delay to give pathologists more time to make the adjustment.

"When we wrote the protocols, we had no idea they were going to be employed in the way they have by the ACS Commission on Cancer," says Carolyn Compton, MD, PhD, chair of the CAP Cancer Committee and the CAP’s representative to the CoC. "But pathologic data is so key to good cancer care that the commission wanted to make these guidelines mandatory immediately."

"We argued successfully that pathologists needed to have time to become conversant with the protocols and implement them in their own practices before they became mandatory."

She adds that the CoC understands the protocols were written originally as "optimal" reporting guidelines, not minimal reporting standards, and that CAP’s Cancer Committee has been working feverishly in the past six months to tailor the protocols to their new role.

"What we’ve done is update and reevaluate every single protocol we’ve published to make certain they reflect the newest classification of tumors and staging systems," Dr. Compton says. "At the same time we’ve subclassified the reporting elements into those that are completely validated and used routinely for patient care—and they would be the ones to be considered mandatory by the commission—and those that are exempt reporting ’suggestions’ in the protocols."

The latter are all data-driven and predictively important, but not yet proven in large studies that definitively show their validity. "We’ve designated them with asterisks, and those elements will be considered optional, not mandatory. So this should quell any arguments that the CoC will be demanding nonessential elements," she says.

The CAP Cancer Committee also feels the mandate is positive in that it will give pathology more say in how cancer reporting is conducted. In fact, says Dr. Compton, who chairs the pathology department at McGill University in Montreal, Quebec, previous pathology reporting standards used by the Commission on Cancer were more or less derived by surgeons, not by pathologists.

"What we’re doing is taking the all-inclusive lists, and pulling from those lists the essentials, the ’must-haves,’ and saying every pathology report has to have those details,’ says Dina Mody, MD, who is an advisor to the Cancer Committee and chair of CAP’s Cytopathology Committee. "If you decide to put in other details, that’s your choice, but the reports are going to be inspected on the ’essential elements.’"

No make-or-break standard

The word "accreditation" may strike fear into the hearts of many. But unlike CAP accreditation of laboratories (which certifies whether laboratories meet federal CLIA ’88 requirements), CoC’s designation of approved cancer programs is itself voluntary and not linked to government regulations that might make a hospital ineligible for reimbursement. M. Asa Carter, manager of the accreditation and standards section of ACS, also stresses that no hospital or medical center would ever fail an accreditation survey on the basis of a single standard such as the protocol mandate.

"We understood that the cancer protocols were intended to be voluntary," Carter says. But she notes that many similar guidelines prepared by other organizations were not developed to be required, and "The Committee on Approvals of the Commission on Cancer felt [cancer reporting] was an important enough issue that it was time to take a stand."

Pathology reports from 2003 will not be examined until 2004, Carter says. And if some of the essential data elements are found to be missing, an amended report can be filed.

Initially, the CoC Committee on Approvals considered requiring the two major protocols—those for colon and breast cancer—but it voted after much discussion to require use of all CAP protocols. After approval by the CAP’s Council on Scientific Affairs and the Board of Governors, the revised cancer protocols will be posted on the CAP Web site, many of them by December. Any protocol not revised by then may be taken out of circulation, and the CoC has said it will not try to mandate anything the College has not approved.

The protocols have been driven primarily by what is needed by the American Joint Committee on Cancer and the International Union Against Cancer, says CAP Cancer Committee member Philip Branton, MD, a staff pathologist at Inova Fairfax Hospital in Fairfax, Va. These groups promulgate the TNM (tumor, node, metastasis) staging system, and their cancer staging manual is used not only by pathologists but also by clinicians.

"When a patient is staged for cancer, the surgical pathology findings are simply one set of the data points used," Dr. Branton says, "You also need to look at radiographic studies, CT scans, nuclear medicine studies, in some cases serum markers for different types of tumors. So all of those data elements need to be considered."

It hasn’t really been demand from the surgical community that has brought the protocol mandate, confirms University of Kentucky Chand-ler Medical Center surgeon Daniel E. Kenady, MD, a member of the CoC and chair of the ad hoc committee to revise standards. "The reporting elements from CAP, we think, are very important just because we are mandating that cancer cases be staged, and we feel really that the most accurate stagings are the ones that utilize the templates," he adds, noting that at his own institution templates have significantly improved accuracy of staging.

"Obviously the pathologist doesn’t have access to the whole patient," says Dr. Kenady, "but the copies of the path reports to the physicians who are responsible for completing the staging forms are used extensively and are much more helpful than operative notes and other parts of the medical record."

Qualms about standardization

The protocols should make life easier for pathologists, Dr. Compton says. "It can be difficult to keep abreast of developments in fields that are rapidly moving like breast and prostate cancer. The protocols are a way that any surgical pathologist can keep up with the most important and essential issues in pathologic analysis without having to read all of the literature themselves."

But it’s what "standardization" connotes that has unsettled some members of the pathology community. Thomas P. Anderson, MD, president of Consulting Pathologists of Connecticut, P.C., is one of them. At his 200-bed community hospital, the four pathologists now use one (non-CAP) template for carcinoma of the breast, but adopting all the protocols will be a huge task.

"I’m not terribly in favor of protocols in the first place because they may come back to bite you in a legal situation. I’ve been practicing pathology for 20 years and am familiar with the types of information the oncologists and other clinicians need. I work very closely with them, and if by any chance a piece of information is not included in the report, they can call me up and say, ’Tom, what about this?’ and I can either tell them or review the case."

Dr. Anderson says the protocols should remain voluntary. "From my point of view, as your day-to-day practicing CAP member, slugging it out at a community hospital without residents, the protocols are an unnecessary burden. Academic pathologists, who primarily devised the protocols, are a different breed with a totally different agenda."

Roger A. Reichert, MD, PhD, chief of the Department of Pathology at St. Luke’s Hospital in St. Louis, Mo., was also an early critic of the CoC initiative to make the protocols mandatory. "CoC was concerned that important staging information was not always present in pathology reports, but commandeering the highly detailed CAP cancer reporting protocols that were intended to be used only as educational tools was the wrong approach to addressing this problem," he says.

In a letter Dr. Reichert wrote to CAP TODAY earlier this year, he used the colon cancer protocol to highlight how some of the items in the original versions of the protocols represent information that is not essential for patient treatment. He also noted that the 90 percent compliance standard set by the CoC is too rigorous to be used in conjunction with these protocols. However, he says the CAP has been responsive to his concerns and he is awaiting the final version of the revised protocols, which his group plans to use as a framework for narrative reports. An opponent of man-datory checklists, Dr. Reichert cited a study that found that synoptic reporting actually reduced physicians’ comprehension of the critical information pathologists were trying to convey (Arch Pathol Lab Med. 2000; 124: 1040-1046).

Uniformity’s benefits

The need for more uniformity in reporting has become increasingly apparent, says Linda Trapkin, DO, a member of the CAP’s Council on Public Affairs and a staff pathologist at St. Joseph’s Hospital in Syracuse, NY. "Twenty years ago, your clinicians got used to what you’d call ’moderately differentiated,’ and everybody was happy. Now a person treated in rural Maine expects the same analysis for treatment options as in New York City, and that can only be done by uniformity of reporting. Those who disagree and say I’m not going with the mainstream, I think they’re wrong."

But there is room for individual approaches, she says. "I probably follow prostate protocols 100 percent, but the one on colon cancer I don’t. There is something on tumor border and lymphocytic invasion. I was not taught to report these aspects and have not seen either as a consensus recommendation in the literature. I have never put either of these descriptions in my reports, and I’ve never been called by the surgeon or the oncologist to do so."

She questions other areas as well. "If you look at the protocols for breast and colon cancer, the instructions when doing lymph node dissection say the entire lymph node is to be put through if it looks benign. I think that’s controversial because most pathologists aren’t doing it yet. I’ve tried it and it’s a lot more work. So pathologists do have the right to look to the CAP to upgrade the CPT payment code for lymph node dissection." Most important, she says, "I think this aspect of the protocols needs to be settled because it is one of the few things that can’t be retrospectively reevaluated, like looking back at a slide to count mitoses."

The breast cancer protocol has run into the most controversy—wrongly, she believes. "I spoke at the House of Delegates meeting, and people from my own delegation said they don’t use the recommended grading system. But if you pick up any recent authoritative textbook, that’s the system being used."

To a certain degree, differences in philosophy and practical needs between those in academia and those in the field are to blame, Dr. Trapkin contends. "The proof that lymphocytic invasion of colon cancer is important is obviously coming from academic centers where research studies are looking at numerous variables, maybe publishing their findings, and the authors may think they are on to something very important, but we in the community hospitals are not hearing feedback from our oncologists that such data is important for diagnostic, prognostic, or treatment decisions. And this does not worry me; if something like lymphocytic infiltration becomes important, it’s still on the slide for retrospective evaluation."

But the trend throughout medicine has been to move toward evidence-based practices. "Whether it’s the drugs clinicians used for heart disease, or antibiotics for infections, 25 years ago it was very idiosyncratic across the country, and sometimes what people were doing was relatively ineffective. Then there was a push to standardize protocols, and our clinical colleagues were presented with algorithms, practice protocols, and clinical guidelines, and they’ve been dealing with them for years longer than we have," Dr. Trapkin points out.

"In the beginning they may also have balked and said this is cookbook medicine, but they’ve calmed on this and now they expect and look for the newest protocol. In pathology we’re still a little cranky and independent."

Once there is agreement on the essentials, Dr. Trapkin expects most pathologists will go along with the protocols, though some will resist changing their style of reporting. The CoC has agreed, however, it doesn’t have to be in tabular form. "So then it becomes whether we are arguing format over content. And I strongly believe that pathologists in this day and age need to be selling themselves lest some commercial outfit comes in and takes over their services. If I can convince pathologists to look at this as a marketing tool, hopefully their attitude will change and they will see that complete and easily understood reports are a marketing plus."

"I’ve lost the vast majority of my prostate business to commercial companies that realize they can make a lot of money off of global billing," she continues. "And when my urologists get a report from a commercial anatomic pathology lab, I’m sure all the recommended information is there, plus the photograph." With office business up for grabs, the urologist is now in the driver’s seat and can say, "This is what I want in my report."

Says the ACS’ Carter: "We don’t say they’re required to use a checklist. We’re asking them to include the data items in their pathology report, and if they choose to do it in a narrative, that’s their choice. However, it’s certainly easier to monitor if they use a checklist."

The ACS Commission on Cancer, she adds, has been told by CAP members that adopting a checklist has saved pathologists a lot of time spent going back to pull slides because they had left something off. "Certainly if I were a physician I would rather have this information up front rather than get it later. If I were a physician I would think I’m placing an additional burden on the pathologist to go back to review slides."

Liability hazards

The resistance by some pathologists is understandable, Dr. Compton acknowledges. "People don’t like to be told what to do, even by their peers." In pathology reporting, standardization has not been the tradition; rather, the tradition has been one of "poetic license," with people being able to say what they wanted on their own reports.

"But we can no longer afford poetic license," she contends. Given the global nature of cancer care, with patients traveling from one center to another for care, for followup, or for clinical trials of new therapies, "it’s absolutely mandatory we all be speaking the same language, and using the same data elements and the same parameters to evaluate those elements."

Making anything mandatory is inevitably going to be controversial, says David Frishberg, MD, chief of pathology at Sinai Hospital of Baltimore in Maryland and a member of the CAP’s Surgical Pathology Committee. But he favors more standardization. "I don’t think patient reports are really designed to be creative exercises. They’re designed to convey information, so I’ve always been strongly for protocols and strongly for a synoptic format on cancer reports." He trained, in fact, in a program where all of the cancer reports were done with protocols and synoptic-style reports, and it led to fewer phone calls from physicians.

When he reviewed a draft of a CAP cancer protocol, however, he admits to getting into a "fairly vibrant disagreement over one of the elements for melanoma, specifically radial versus vertical growth phase. I felt that should be optional; it’s vaguely and variably defined in the literature, so it’s hard to expect general pathologists to apply it, and not everyone accepts it as important."

"But the biggest problem with mandating specific elements—even ones for which there is a relatively broad consensus—is that it’s a moving target. The literature changes, and things considered important now could become less so in the future and vice versa," he says.

Many cite another possible hazard: legal liability. Potentially, the protocols do increase pathologists’ exposure, Dr. Frishberg acknowledges. "I think if we have mandatory inclusion of elements that even experts might differ on, a pathologist who declined to put it in could be [unfairly] accused of violating standards of practice. But I would say for a lot of the elements, it probably would be a violation of standards of practice" to leave them out. Furthermore, he says, the greater the number of specific items listed, "the more there are that can be challenged as ’wrong’ in a medicolegal action."

Even well-meaning educational efforts carry this risk, Dr. Mody cautions. "Unfortunately, the legal profession has mucked it up wherever they’ve put their finger. If you use the word ’guideline’ or ’protocol,’ the next thing you know if the patient has a bad outcome and somebody did not include an obscure detail from the guideline, there is a possibility some lawyer may imply you were negligent."

In cytology, she notes, Pap test lookback was an educational and quality improvement activity. "Guess what? It became part of the regulations. Now, every time there is a patient with cancer, that educational and quality assurance activity has become a bonanza for trial lawyers."

But following protocols can be a safeguard in litigation, says M. Elizabeth Hammond, MD, chair of the Department of Pathology at LDS Hospital/Intermountain Healthcare in Salt Lake City. "In Intermountain Healthcare, we try to develop protocols for many aspects of care including pathology reporting. Our legal counsel has noted that they provide a potent weapon in litigation." If a protocol has been adopted and followed, she says, it becomes the responsibility of the lawyer for the injured party to show how the protocol, rather than the individual physician using the protocol, might have produced harm to the patient. "Cases where this defense has been employed have been ruled on favorably by judges and prelitigation panels," she says.

Accentuating the essentials

Whether a litigation help or hindrance, the protocols are needed, contends Dr. Mody, of Baylor College of Medicine in Houston. "Unfortunately, I still see reports where the tumor size is not reported when it’s required. Sometimes I get reports from rural pathologists that are excellent and have all the data points, and I could get a report from within a short radius of the Texas medical center, the largest medical center in the world, and not have those."

"You can’t say the little guy out in the community is not doing a good job," Dr. Mody continues. "It’s just human nature. If you’re doing a rambling report, maybe you just forget. But it’s nice to have nine or 10 things on a checklist. If you forgot it in the narrative, well, you can check it off at the bottom."

Says Dr. Branton: "There are a lot of reports floating around that leave out a lot of critical data elements. Some people don’t necessarily keep up with the times. Maybe they’re in smaller practices, either rural or in-city, or maybe they’re simply in a practice setting that rarely sees a breast biopsy—for example, a VA hospital. If you haven’t looked at a breast cancer for four or five years, you might not be quite up to snuff. With the protocols, a working pathologist in a smaller group setting has a convenient, concise, one-stop shopping format for what needs to be put into a report."

He says the synoptic format averts the problem of critical data being "buried in the verbiage" of a traditional report. "In residency we were all taught the Dickensian narrative, to give very lengthy microscopic descriptions of everything. Part of the reasoning for that was very sound, and on the [CAP] committee the resistance we’ve encountered in the pathology community has been a concern that if you have a template kind of pathology report, it’s all cookie-cutter; we’re all the same and there’s no role for individual skill. But the bottom line is we still need a highly trained and skilled person looking at the slide. The difference is the reports are going to be stripped down and streamlined so that oncologists and surgeons can look at them and quickly gather the information they need."

For any cancer section, "the kind of detail needed by the treating clinician has just exploded in the last five years," he adds. "When I was a resident 15 years ago, they wanted to know how big the breast tumor was, was the margin involved, were the lymph nodes involved, and was there any pre-invasive cancer or duct car-ci-no-ma in situ. Now a report has 15 to 18 parameters." For example, multi-cen-ter studies have found that the pres-ence of vascular invasion is pre-dic-tive of a worse course. "Those patients would probably be offered che-mo-therapy even if the lymph nodes don’t have tumors in them. Fifteen years ago we didn’t know that."

In the 1950s through the 1970s, he says, "surgical pathology was fairly static. It was basically light microscopy, looking at slides and saying this is or isn’t cancer and how big it is. People of that era are sometimes a little bewildered by what’s going on, and that’s one reason why CAP has been working on these cancer protocols."

"Any good pathologist would not want this to happen," says Dr. Compton, "but with the high volume of work and the labor intensity of cancer cases, some reports get signed out inadvertently with missing information needed for patient management." The national cancer database bears out this fact, she adds. "It’s clear, for instance, that as many as one quarter of all colon cancer cases get signed out without a lymph node assessment, and as many as 50 percent of all sarcoma cases get signed out without a grade."

Advice and consent

Dr. Hammond became involved in the cancer protocols in the early 1990s, later chairing the Cancer Committee. But the committee started convening task forces on multiple specific tumor types as far back as the mid-80s, beginning with Hodgkin’s disease, breast cancer, and colon cancer. "In every case," she says, "the same practice was followed. A panel of experts including clinicians and pathologists would meet and talk together about what information was important and how it should be reported so that it was clear. Then they came up with a protocol and published it in the Archives."

None of the task forces’ work was easy, she adds. "I think all of them have been difficult. With prostate cancer there was difficulty getting unanimity of opinion among the experts, and the same with bladder cancer." The key to their success, however, was that clinicians drove the conversation. "We tried to pick people who were reasonable human beings. There were surgeons as well as radiation and medical oncologists, and we said, ’What do you need and how should we present it so you can use it?’"

While the ACS was not directly involved, "it was the organization that really sponsored the early efforts to stage cancers in the U.S., and they wanted staging information to be improved, so they watched with interest as the efforts by the Cancer Committee were going on."

"What’s being mandated are really just the elements," Dr. Hammond emphasizes. "Not how the pathologist gives the information but merely that he or she does. For example, in breast cancer you give the lymph node status; in prostate cancer you give the Gleason score. But how you write that in the report is not part of the requirements."

At LDS Hospital/Intermountain Healthcare, the checklist format has been used exclusively for cancer reporting since 1992—largely because of a review she performed of 1991 pathology reports (Arch Pathol Lab Med. 1997;121:1171-1175). "I found that only 299 of 356 narrative breast cancer reports were complete and clear. Thirty-two reports contained missing information, such as size of the tumor or margin status, and another 25 were confusing because of the way the information was reported in the narrative. A log of phone calls showed the majority of calls by oncologists were to report missing information. The large number of phone calls on cancer reports saying ’What do you mean?’ caused us to change to a synoptic format."

When the oncologists were surveyed later to see if they liked the synoptic format, they answered that they would never go back to the previous system, because now they knew every single time what grading scheme was being used, and there was no confusion from one doctor to another. "Another valuable aspect of synoptic or checklist reporting is that it allows these discrete data elements to be captured in a data warehouse to help in cancer care modifications," Dr. Hammond says. To make it possible to search the data, some degree of standardization of terminology or format or both is necessary—even more so now that so many institutions are moving to computerized medical records.

"I think the place for narrative is dwindling," Dr. Hammond continues. "The reason is that clinicians are forced to work with less and less time. If you can read something as a bulleted list, you’re much more likely to get to that kernel of information more quickly than if you have to read an entire report. That’s why doctors like to read abstracts instead of whole manuscripts, so I think the day of writing long narrative reports is going to disappear." Very few of the reports produced by successful clinical laboratories are done in narrative format anymore, she adds, except when a comment is needed to clarify difficulties with the checklist format.

For Joel Sennesh, MD, vice chairman of the Department of Pathology at Inova Fairfax Hospital in Fairfax, Va., there will not be much of an adjustment to make since his tertiary-level care center has been using the protocols for several years. "I’ve heard anecdotally that the CoC wants to give its accreditation to hospitals that use this systematized reporting. But we try to do as many of the common cancer diagnoses as possible in protocol format, since our surgeons and oncologists have indicated they’d like to see reports as giving a TNM." In addition, most of the reports he sees are from academic centers like Duke or Johns Hopkins universities, "which all seem to be adhering to those protocols, so we really don’t see that much that diverges from the protocols."

In his view, cancer reporting should be a straightforward matter. "I see my job as being somebody who helps compile objective data. For the benefit of future science and future patients, it’s probably worthwhile to be as standardized as we can."

Are the mandatory protocols a sign that surgeons are decreeing medical practices to pathologists? Dr. Compton says the reverse is true. "Actually, pathologists were dictated to by surgeons more before they got these guidelines. Now at least the guidelines are pathology-generated, and they’ve gone through a huge gauntlet of pathologists at all levels, and we’ve taken feedback from all CAP members and incorporated it into these guidelines."

Says Dr. Branton: "I think CAP is doing a superb job by showing leadership in this area. Certainly from the reports I’ve gotten, the surgeons are absolutely delighted with what CAP is doing, and they’ve seized on the protocols as exactly what they’re looking for."

It’s important to remember that College members developed these protocols, Dr. Frishberg notes. "We’re very lucky we didn’t have non-pathologists specifying protocols to pathologists. Other entities [like the Association of Directors of Anatomic and Surgical Pathology] have published their protocols, and it’s good that the College really stepped up to the bat to take responsibility here. CoC is obviously the authority for accrediting cancer programs, so if they had to pick someone to do it, I’m glad it’s us."

Anne Paxton is a writer in Seattle. Each of CAP’s approved cancer protocols consists of a long form, containing all information for staging, grading, and prognostic factor analysis, and a short form, or checklist. The 2002 versions are available free on the Web at www.cap.org.